Performance issue when calling java module

Hi experts,
It takes much time to call Java module 'JBND_CRM_SVY_XSLT_TRANSFORM' from abap function module.
When I executed runtime analysis via t-cd:se30,
most of time is consumed in 'wait for RFC' .
When the java module is called, it always takes much time.
I think VMC memory management is not the cause,
because the issue also occurs at first call.
The kernel patch level is 146 of release 700.
Do you have any idea to improve performance?
Regards,
Fukuhara Yohei

Hi
try with this code
remove
READ TABLE i_events INTO wa_events WITH KEY name = 'USER_COMMAND'.
IF sy-subrc EQ 0.
wa_events-form = 'USER_COMMAND'.
MODIFY i_events FROM wa_events TRANSPORTING form WHERE name =
wa_events-name.
ENDIF.
and change alv
CALL FUNCTION 'REUSE_ALV_GRID_DISPLAY'
EXPORTING
i_callback_program = sy-repid
i_callback_user_command = 'USER_COMMAND'
it_fieldcat = i_ztal_prof1_cat[]
i_save = 'A'
TABLES
t_outtab = i_ztal_prof1.
Marco

Similar Messages

Possible thread issues when calling java from C++/Delphi/C#

Hi,
We have a java API for one of our applications. This API has wrapper API's in C++, Delphi and C#. There are two main top level classes in te API - Client and Receiver. Both haave a receive() method.
We used to ship JRE1.3 with the API installation - this was needed to suport the user applications. Recently we haev moved to java 1.4 as the least suported version.
We have two ways of using the java API in our wrapper API's.
Method one, we instantiate the client, and call client.receive(). There are no threads etc involved here. This is working fine.
In the second method, we set a callback on the client, which starts a thread (in C++/Delphi or C#) which calls receiver.receive() and tries to get the message. This is failing.
It fails at a line which simply says "DocumentBuilderFactory.newInstance()".
This call has already been made when connecting the client and it is successful because we can see the client connected.
Whether we do client.receive() or receiver.receive() in a thread, the code passes through this line.But is throwing the exception only when we come to it through a thread.
Any ideas on what I need to watch out for when calling methods on a JVM from a thread in C++/Delphi or C#?
Thanks
Hari
PS: I know we are using some really old versions, but thats needed and out of my control.

Hi,
I managed to solve this - finally!
This is what was the cause in our case - should apply to all situations though.
We had the Xerces classes packaged in a custom jar file along with other XML stuff we have. And this mega-jar was on the class path.
We were running everything on Java 1.3 and it worked fine.
But Java 1.4 wrecked all havoc.
So here is what I did:
1. I tried explicitly doing Class.forName() with the name of the Xerces class that was not found. I did this at the same point in the code where the exception was being thrown.
2. The class was successfully loaded indicating that it was on the class path somewhere.
3. Java 1.4 made some changes in terms of class loaders - so suspecting that, I tried printing out the System and the Context class loaders at the point where the exception was being thrown - the system class loader was a valid object, but context class loader was null
4. We have another point in the code base where the exact same lines were used, and these are getting successfully executed even in the multi threaded situation - this is the start up sequence.
5. I printed the two class loaders at this point - they are both valid objects, no nulls.
With this, I narrowed down the scope and was convinced that the class loaders are the cause.
A search on Google and after looking the the source code for Xerces, we realized that the Xerces API uses the Context Class Loader to resolve and load the DocumentFactoryImpl class - this change is from Java 1.4 onwards.
In our case, the JNI AttachCurrentThread method called from C++/C# and Delphi was attaching the thread to the JVM, but each time the context class loader was null.
We could not find anything on how we could make the method to put a non null context classloader.
So, we changed all our JNI interfaces to do the following
//attach the current thred
jvm->AttachCurrentThread(...)
//now get the system class loader and set it as the context class loader on the current thread
//get things from the Thread class
jclass thread = get the thread class
jmethod currentThread = get the static method currentThread()
jobject thread_object = get the current thread object into this
jmethod setContextClassLoader = get the setContextClassLoader() method on the thread
//get the SystemClassLoader
jclass clsLoader = get the class for ClassLoader
jmethodid getSysClsLoader = get the method ID for the getSystemClassLoader() method
jobect sysClsLoader = get the system class loader by calling the above method
//set the context class loader on the thread by calling set context class loader with the sys class loader as the argumentUnfortunately I don't have the code with me right now,otherwise would have shared the same.

Browser performance issue when Calling BI Publisher report from OBIEE dashb

Hi Techies,
We installed OBIEE 10.1.3.3 on SunOS 5.9 unix box.This is PROD env.We are calling BIPublisher report from OBIEE dashboard by clicking the BIP reports link on OBIEE dashboard.After clikcing the link BIPublisher report is opening in another browser window.But is it taking time.
Plz tell me how can we reduce the time taken to open BIPublisher report window when we cilck on the report name in OBIEE dashboard
Thanks,

Suresh
There is no out of the box method to get the EBS report to OBIEE. You could look into moving the EBS report to a mountable web server directory after it completes and then show it in OBIEE via a link or html object. You will need to rename the report of course, every time you run it its going to change names.
In an after report trigger you could call a small java class to move and rename the file for you.
Alternatively, recreate the report in the standalone version. Be aware that you are going to have to set orgs yourself, BIP is not aware of orgs outside of EBS
Tim

Error BT616 when calling function module SXPG_COMMAND_EXECUTE in background

Hi All,
We use function module SXPG_COMMAND_EXECUTE with a custom command we defined in SM69 to move files in unix (mv command).
The function module call has worked fine for almost a year and recently we have been seeing an error (BT616) in our job lob (SM37) when the program is run in background. We have not been able to reproduce the error in foreground mode and it seems to be occuring only periodically in the background. (The appropriate SAP authorization objects where assigned to the batch job ID and the steps on the batch job.) We are in the process of setting up the trace flag and performing analysis on the trace log via ST11 to help identify the issue.
After perform analysis on SXPG_COMMAND_EXECUTE, the error is occurring when calling function module SAPXPG_END_XPG for exception 2, system failure, yet function module SAPXPG_END_XPG does not exist. I assume this is a program at the operating system level and is just a signature of the parameters to be passed to the operating system program.
Below is part of the SAP function module SXPG_COMMAND_EXECUTE that is failing.
* Now we have to wait for the termination of the external
* command if the caller wants us to.
    IF TERMINATIONWAIT = 'X'.
      CALL FUNCTION 'SAPXPG_END_XPG'
        DESTINATION DESTINATION
        IMPORTING   EXITSTAT = STATUS
                    EXITCODE = EXITCODE
        TABLES      LOG      = LOG
        EXCEPTIONS COMMUNICATION_FAILURE = 1 MESSAGE MSG
                    SYSTEM_FAILURE        = 2 MESSAGE MSG.
I performed a where used on function module SXPG_COMMAND_EXECUTE, and most of SAP programs call the function module with the parameter TERMINATIONWAIT = 'X', so I assume we should pass X as well.
Any ideas on what could be causing this issue?
Mike Vondran

I also remember I have this kind of issue, as I have some UNIX script at OS( UNIX) level . The problem was with the ID , as it dont have proper authorization at OS level ( UNIX ) . Please check this ID authorization. This could be the one of reasons if youre sure from SAP standpoint.
Hope thisll give you some guide line..
Thanks
Bye

Error Report when calling Java thru C

Currently we have encountered some issues with our C code calling java module in NT- Itanium 64bit machine. We get the following error message (we don't get this on the 32 bit):
Another exception has been detected while we were handling last error.
Dumping information about last error:
ERROR REPORT FILE = (N/A)
PC = 0x0000000077f69684
SIGNAL = -1073741819
FUNCTION NAME = (N/A)
OFFSET = 0xFFFFFFFF
LIBRARY NAME = (N/A)
Please check ERROR REPORT FILE for further information, if there is any.
Good bye.
Has anybody seen this and has any suggestions? using JDK 1.4
Thanks

I was thinking JVM should the classpath when it is invoked. But obviously it is not and we need to assign the classpath ourselves when invoking from native side. Thanks for pointing that out. For other people's benefit the following is the code piece I for initializing JVM.
char* str1;
          char* str2;
          char* classpath;
          str1 = "-Djava.class.path=";
          str2 = std::getenv("CLASSPATH");
          classpath = (char*)malloc(strlen(str1) + strlen(str2) + 1);
          strcpy(classpath, str1);
          strcat(classpath, str2);
          if (str2 == NULL)     {
               printf("CLASSPATH environment variable is not defined.");
               exit(1);
          m_VmArgs.version = JNI_VERSION_1_4;
          m_options[0].optionString= classpath;
          m_VmArgs.options = m_options;
          m_VmArgs.nOptions = 1;
          m_VmArgs.ignoreUnrecognized = JNI_TRUE;
          JNI_GetDefaultJavaVMInitArgs(&m_VmArgs);
          /* Create the Java VM */
          m_Res = JNI_CreateJavaVM(&m_Jvm, (void**)&m_Env, &m_VmArgs);
          if (m_Res < 0) {
               printf("Can't create JVM\n");
               exit(1);
thanks,
chancellor

Performance issue for this function-module(HR_TIM_REPORT_ABSENCE_DATA)

Hi Friends
I am having performance issue for this function-module(HR_TIM_REPORT_ABSENCE_DATA) and one my client got over 8 thousend employees . This function-module taking forever to read the data. is there any other function-module to read the absences data IT2001 .
I did use like this .if i take out this F.M 'HR_TIM_REPORT_ABSENCE_DATA_INI' its not working other Function-module.please Suggest me .
call function 'HR_TIM_REPORT_ABSENCE_DATA_INI'
exporting "Publishing to global memory
option_string = option_s "string of sel org fields
trig_string = trig_s "string of req data
alemp_flag = sw_alemp "all employee req
infot_flag = space "split per IT neccessary
sel_modus = sw_apa
importing
org_num = fdpos_lines "number of sel org fields
tables
fieldtab = fdtab "all org fields
field_sel = fieldnametab_m. "sel org fields
To Read all infotypes from Absences type.
RP_READ_ALL_TIME_ITY PN-BEGDA PN-ENDDA.
central function unit to provide internal tables: abse orgs empl
call function 'HR_TIM_REPORT_ABSENCE_DATA'
exporting
pernr = pernr-pernr
begda = pn-begda
endda = pn-endda
IMPORTING
SUBRC = SUBRC_RTA
tables
absences = absences_01
org_fields = orgs
emp_fields = empl
REFTAB =
APLTAB =
awart_sel_p = awart_s[]
awart_sel_a = awart_s[]
abstp_sel = abstp_s[]
i0000 = p0000
i0001 = p0001
i0002 = p0002
i0007 = p0007
i2001 = p2001
i2002 = p2002
i2003 = p2003.
Thanks & Regards
Reddy

guessing will not help you much, check with SE30 to get a better insight
SE30
The ABAP Runtime Trace (SE30) - Quick and Easy
what is the total time, what are the Top 10 in the hitlist.
Siegfried

Oracle Retail 13 - Performance issues when open, save, approving worksheets

Hi Guys,
Recently we started facing performance issues when we started working with Oracle Retail 13 worksheets from within the java GUI at clients desktops.
We run Oracle Retail 13.1 powered by Oracle Database 11g R1 and AS 10g in latest release.
Issues:
- Opening, saving, approving worksheets with approx 9 thousands of items takes up to 15 minutes.
- Time for smaller worksheets is also around 10 minutes just to open a worksheet
- Also just to open multiple worksheets takes "ages" up to 10-15 minuts
Questions:
- Is it expected performance for such worksheets?
- What is your experience with Oracle Retail 13 in terms of performance while working with worksheets - how much time does it normally take to open edit save a worksheet?
- What are the average expected times for such operations?
Any feedback and hints would be much appreciated.
Cheers!!

Hi,
I guess you mean Order/Buyer worksheets?
This is not normal, should be quicker, matter of seconds to at most a minute.
Database side tuning is where I would look for clues.
And the obvious question: remember any changes to anything that may have caused the issue? Are the table and index statistics freshly gathered?
Best regards, Erik Ykema

Performance issues when creating a Report / Query in Discoverer

Hi forum,
Hope you are can help, it involves a performance issues when creating a Report / Query.
I have a Discoverer Report that currently takes less than 5 seconds to run. After I add a condition to bring back Batch Status that = Posted we cancelled the query after reaching 20 minutes as this is way too long. If I remove the condition the query time goes back to less than 5 seconds.
Please see attached the SQL Inspector Plan:
Before Condition
SELECT STATEMENT
SORT GROUP BY
VIEW SYS
SORT GROUP BY
NESTED LOOPS OUTER
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
TABLE ACCESS BY INDEX ROWID GL.GL_CODE_COMBINATIONS
AND-EQUAL
INDEX RANGE SCAN GL.GL_CODE_COMBINATIONS_N2
INDEX RANGE SCAN GL.GL_CODE_COMBINATIONS_N1
TABLE ACCESS BY INDEX ROWID APPLSYS.FND_FLEX_VALUES
INDEX RANGE SCAN APPLSYS.FND_FLEX_VALUES_N1
TABLE ACCESS BY INDEX ROWID APPLSYS.FND_FLEX_VALUE_SETS
INDEX UNIQUE SCAN APPLSYS.FND_FLEX_VALUE_SETS_U1
TABLE ACCESS BY INDEX ROWID APPLSYS.FND_FLEX_VALUES_TL
INDEX UNIQUE SCAN APPLSYS.FND_FLEX_VALUES_TL_U1
INDEX RANGE SCAN APPLSYS.FND_FLEX_VALUE_NORM_HIER_U1
TABLE ACCESS BY INDEX ROWID GL.GL_JE_LINES
INDEX RANGE SCAN GL.GL_JE_LINES_N1
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
INDEX UNIQUE SCAN GL.GL_SETS_OF_BOOKS_U2
TABLE ACCESS BY INDEX ROWID GL.GL_JE_HEADERS
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
INDEX UNIQUE SCAN GL.GL_DAILY_CONVERSION_TYPES_U1
TABLE ACCESS BY INDEX ROWID GL.GL_JE_SOURCES_TL
INDEX UNIQUE SCAN GL.GL_JE_SOURCES_TL_U1
INDEX UNIQUE SCAN GL.GL_JE_CATEGORIES_TL_U1
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
INDEX UNIQUE SCAN GL.GL_JE_BATCHES_U1
INDEX UNIQUE SCAN GL.GL_BUDGET_VERSIONS_U1
INDEX UNIQUE SCAN GL.GL_ENCUMBRANCE_TYPES_U1
INDEX UNIQUE SCAN GL.GL_SETS_OF_BOOKS_U2
TABLE ACCESS BY INDEX ROWID GL.GL_JE_BATCHES
INDEX UNIQUE SCAN GL.GL_JE_BATCHES_U1
INDEX UNIQUE SCAN GL.GL_SETS_OF_BOOKS_U2
INDEX UNIQUE SCAN GL.GL_JE_BATCHES_U1
TABLE ACCESS BY INDEX ROWID GL.GL_PERIODS
INDEX RANGE SCAN GL.GL_PERIODS_U1
After Condition
SELECT STATEMENT
SORT GROUP BY
VIEW SYS
SORT GROUP BY
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS
NESTED LOOPS
NESTED LOOPS OUTER
NESTED LOOPS
TABLE ACCESS FULL GL.GL_JE_BATCHES
INDEX UNIQUE SCAN GL.GL_SETS_OF_BOOKS_U2
INDEX UNIQUE SCAN GL.GL_JE_BATCHES_U1
TABLE ACCESS BY INDEX ROWID GL.GL_JE_HEADERS
INDEX RANGE SCAN GL.GL_JE_HEADERS_N1
INDEX UNIQUE SCAN GL.GL_SETS_OF_BOOKS_U2
INDEX UNIQUE SCAN GL.GL_ENCUMBRANCE_TYPES_U1
INDEX UNIQUE SCAN GL.GL_DAILY_CONVERSION_TYPES_U1
INDEX UNIQUE SCAN GL.GL_BUDGET_VERSIONS_U1
TABLE ACCESS BY INDEX ROWID GL.GL_JE_SOURCES_TL
INDEX UNIQUE SCAN GL.GL_JE_SOURCES_TL_U1
INDEX UNIQUE SCAN GL.GL_JE_CATEGORIES_TL_U1
INDEX UNIQUE SCAN GL.GL_JE_BATCHES_U1
TABLE ACCESS BY INDEX ROWID GL.GL_JE_LINES
INDEX RANGE SCAN GL.GL_JE_LINES_U1
INDEX UNIQUE SCAN GL.GL_SETS_OF_BOOKS_U2
TABLE ACCESS BY INDEX ROWID GL.GL_CODE_COMBINATIONS
INDEX UNIQUE SCAN GL.GL_CODE_COMBINATIONS_U1
TABLE ACCESS BY INDEX ROWID GL.GL_PERIODS
INDEX RANGE SCAN GL.GL_PERIODS_U1
TABLE ACCESS BY INDEX ROWID APPLSYS.FND_FLEX_VALUES
INDEX RANGE SCAN APPLSYS.FND_FLEX_VALUES_N1
INDEX RANGE SCAN APPLSYS.FND_FLEX_VALUE_NORM_HIER_U1
TABLE ACCESS BY INDEX ROWID APPLSYS.FND_FLEX_VALUES_TL
INDEX UNIQUE SCAN APPLSYS.FND_FLEX_VALUES_TL_U1
TABLE ACCESS BY INDEX ROWID APPLSYS.FND_FLEX_VALUE_SETS
INDEX UNIQUE SCAN APPLSYS.FND_FLEX_VALUE_SETS_U1
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
INDEX UNIQUE SCAN GL.GL_JE_HEADERS_U1
Is there anything i can do in Discoverer Desktop / Administration to avoid this problem.
Many thanks,
Lance

Hi Rod,
I've tried the condition (Batch Status||'' = 'Posted') as you suggested, but the qeury time is still over 20 mins. To test i changed it to (Batch Status||'' = 'Unposted') and the query was returned within seconds again.
Ive been doing some more digging and have found the database view that is linked to the Journal Batches folder. See below.
I think the problem is with the column using DECODE. When querying the column in TOAD the value of P is returned. But in discoverer the condition is done on the value Posted. Im not too sure how DECODE works, but think this could be the causing some sort of issue with Full Table Scans. How do we get around this?
Lance
DECODE( JOURNAL_BATCH1.STATUS,
'+', 'Unable to validate or create CTA',
'+*', 'Was unable to validate or create CTA',
'-','Invalid or inactive rounding differences account in journal entry',
'-*', 'Modified invalid or inactive rounding differences account in journal entry',
'<', 'Showing sequence assignment failure',
'<*', 'Was showing sequence assignment failure',
'>', 'Showing cutoff rule violation',
'>*', 'Was showing cutoff rule violation',
'A', 'Journal batch failed funds reservation',
'A*', 'Journal batch previously failed funds reservation',
'AU', 'Showing batch with unopened period',
'B', 'Showing batch control total violation',
'B*', 'Was showing batch control total violation',
'BF', 'Showing batch with frozen or inactive budget',
'BU', 'Showing batch with unopened budget year',
'C', 'Showing unopened reporting period',
'C*', 'Was showing unopened reporting period',
'D', 'Selected for posting to an unopened period',
'D*', 'Was selected for posting to an unopened period',
'E', 'Showing no journal entries for this batch',
'E*', 'Was showing no journal entries for this batch',
'EU', 'Showing batch with unopened encumbrance year',
'F', 'Showing unopened reporting encumbrance year',
'F*', 'Was showing unopened reporting encumbrance year',
'G', 'Showing journal entry with invalid or inactive suspense account',
'G*', 'Was showing journal entry with invalid or inactive suspense account',
'H', 'Showing encumbrance journal entry with invalid or inactive reserve account',
'H*', 'Was showing encumbrance journal entry with invalid or inactive reserve account',
'I', 'In the process of being posted',
'J', 'Showing journal control total violation',
'J*', 'Was showing journal control total violation',
'K', 'Showing unbalanced intercompany journal entry',
'K*', 'Was showing unbalanced intercompany journal entry',
'L', 'Showing unbalanced journal entry by account category',
'L*', 'Was showing unbalanced journal entry by account category',
'M', 'Showing multiple problems preventing posting of batch',
'M*', 'Was showing multiple problems preventing posting of batch',
'N', 'Journal produced error during intercompany balance processing',
'N*', 'Journal produced error during intercompany balance processing',
'O', 'Unable to convert amounts into reporting currency',
'O*', 'Was unable to convert amounts into reporting currency',
'P', 'Posted',
'Q', 'Showing untaxed journal entry',
'Q*', 'Was showing untaxed journal entry',
'R', 'Showing unbalanced encumbrance entry without reserve account',
'R*', 'Was showing unbalanced encumbrance entry without reserve account',
'S', 'Already selected for posting',
'T', 'Showing invalid period and conversion information for this batch',
'T*', 'Was showing invalid period and conversion information for this batch',
'U', 'Unposted',
'V', 'Journal batch is unapproved',
'V*', 'Journal batch was unapproved',
'W', 'Showing an encumbrance journal entry with no encumbrance type',
'W*', 'Was showing an encumbrance journal entry with no encumbrance type',
'X', 'Showing an unbalanced journal entry but suspense not allowed',
'X*', 'Was showing an unbalanced journal entry but suspense not allowed',
'Z', 'Showing invalid journal entry lines or no journal entry lines',
'Z*', 'Was showing invalid journal entry lines or no journal entry lines', NULL ),

Error :Required parameters missing when calling up module MARC_SINGLE_READ

Dear all,
When trying to do the production order confirmation i am getting the error "Required parameters missing when calling up module MARC_SINGLE_READ".How can this error be solved?
Thanks,
Kumar

Hi
Please activate the corresponding programs again. I think you got a short dump for this error, if yes, you can see the program's name there.
Best Regards.
Leon.

Performance Issues when editing large PDFs

We are using Adobe 9 and X Professional and are experiencing performance issues when attempting to edit large PDF files. (Windows 7 OS). When editing PDFs that are 200+ pages, we are seeing pregnated pauses (that feel like lockups), slow open times and slow to print issues.
Are there any tips or tricks with regard to working with these large documents that would improve performance?

You said "edit." If you are talking about actual editing, that should be done in the original and a new PDF created. Acrobat is not a very good editing tool and should only be used for minor, critical edits.
If you are talking about simply using the PDF, a lot depends on the structure of the PDF. If it is full of graphics, it will be slow. You can improve this performance by using the PDF Optimize to reduce graphic resolution and such. You may very likely have a bloated PDF that is causing the problem and optimizing the structure should help.
Be sure to work on a copy.

Performance issue when a Direct I/O option is selected

Hello Experts,
One of my customers has a performance issue when a Direct I/O option is selected. Reports that there was increase in memory usage when Direct I/O storage option is selected when compared to Buffered I/O option.
There are two applications on the server of type BSO. When using Buffered I/O, experienced a high level of Read and Write I/O's. Using Direct I/O reduces the Read and Write I/O's, but dramatically increases memory usage.
Other Information -
a) Environment Details
HSS - 9.3.1.0.45, AAS - 9.3.1.0.0.135, Essbase - 9.3.1.2.00 (64-bit)
OS: Microsoft Windows x64 (64-bit) 2003 R2
b) What is the memory usage when Buffered I/O and Direct I/O is used? How about running calculations, database restructures, and database queries? Do these processes take much time for execution?
Application 1: Buffered 700MB, Direct 5GB
Application 2: Buffered 600MB to 1.5GB, Direct 2GB
Calculation times may increase from 15 minutes to 4 hours. Same with restructure.
c) What is the current Database Data cache; Data file cache and Index cache values?
Application 1: Buffered (Index 80MB, Data 400MB), Direct (Index 120MB; Data File 4GB, Data 480MB).
Application 2: Buffered (Index 100MB, Data 300MB), Direct (Index 700MB, Data File 1.5GB, Data 300MB)
d) What is the total size of the ess0000x.pag files and ess0000x.ind files?
Application 1: Page File 20GB, Index 1.7GB.
Application 2: Page 3GB, index 700MB.
Any suggestions on how to improve the performance when Direct I/O is selected? Any performance documents relating to above scenario would be of great help.
Thanks in advance.
Regards,
Sudhir

Sudhir,
Do you work at a help desk or are you a consultant? you ask such a varied range of questions I think the former. If you do work at a help desk, don't you have a next level support that could help you? If you are a consultant, I suggest getting together with another consultant that actually knows more. You might also want to close some of your questions,. You have 24 open and perhaps give points to those that helped you.

How to get around a performance issue when dealing with a lot of data

Hello All,
This is an academic question really, I'm not sure what I'm going to do with my issue, but I have some options. I was wondering if anyone would like to throw in their two cents on what they would do.
I have a report, the users want to see all agreements and all conditions related to the updating of rebates and the affected invoices. From a technical perspective ENT6038-KONV-KONP-KONA-KNA1. THese are the tables I have to hit. The problem is that when they retroactively update rebate conditions they can hit thousands of invoices, which blossoms out to thousands of conditions...you see the problem. I simply have too much data to grab, it times out.
I've tried everything around the code. If you have a better way to get price conditions and agreement numbers off of thousands of invoices, please let me know what that is.
I have a couple of options.
1) Use shared memory to preload the data for the report. This would work, but I'm not going to know what data is needed to be loaded until report run time. They put in a date. I simply can't preload everything. I don't like this option much.
2) Write a function module to do this work. When the user clicks on the button to get this particular data, it will launch the FM in background and e-mail them the results. As you know, the background job won't time out. So far this is my favored option.
Any other ideas?
Oh...nope, BI is not an option, we don't have it. I know, I'm not happy about it. We do have a data warehouse, but the prospect of working with that group makes me whince.

My two cents - firstly totally agree with Derick that its probably a good idea to go back to the business and justify their requirement in regards to reporting and "whether any user can meaningfully process all those results in an aggregate". But having dealt with customers across industries over a long period of time, it would probably be bit fanciful to expect them to change their requirements too much as in my experience neither do they understand (too much) technology nor they want to hear about technical limitations for a system etc. They want what they want if possible yesterday!
So, about dealing with performance issues within ABAP, I'm sure you must be already using efficient programming techniques like using Hash internal tables with Unique Keys, accessing rows of the table using Field-Symbols and all that but what I was going to suggest to you is probably look at using [Extracts|http://help.sap.com/saphelp_nw04/helpdata/en/9f/db9ed135c111d1829f0000e829fbfe/content.htm]. I've had to deal with this couple of times in the past when dealing with massive amount of data and I found it to be very efficient in regards to performance. A good point to remember when using Extracts that, I quote from SAP Help, "The size of an extract dataset is, in principle, unlimited. Extracts larger than 500KB are stored in operating system files. The practical size of an extract is up to 2GB, as long as there is enough space in the filesystem."
Hope this helps,
Cheers,
Sougata.

Sys.UTL_DBWS issue when calling web service .......

Hi,
Background Details :
Linux : 2.6.18-164.10.1.0.1.el5 #1 SMP Fri Jan 8 02:34:10 EST 2010 x86_64 x86_64 x86_64 GNU/Linux
Web Server : Apache Tomcat 6.0.20
Database : Oracle 10.2.0.4
Utility that we use to consume the Web Service from PL/SQL : JPublisher 10g Release 2 (10.2)
Load Balancer : Citrix Netscaler
We are using the Oracle built in package called UTL_DBWS to invoke the Java web service. The web service is been called through a Load Balancer called Citrix Netscaler.
Load balancer has been configured to use the Source IP mechanism to route the request to the appropriate web service.
Problem :
When invoking the web service from Pl/SQL (with the below piece of code) there are multiple times the web servers are been called which is supposed to be only once. The number of times the call varies is about to 2 or 3.
Query :
Hence we need to understand in general how the UTL_DBWS works ? Are there any issues when the call being made through the LoadBalancer ?
The default time out value while connecting ? If we want to override the default timeout how we can do that ?
Code for reference :
BEGIN
l_service := sys.UTL_DBWS.create_service (
wsdl_document_location => sys.URIFACTORY.getURI(l_wsdl_url),
service_name => l_service_qname);
l_call := sys.UTL_DBWS.create_call (
service_handle => l_service,
port_name => l_port_qname,
operation_name => l_operation_qname);
l_input_params(0) := sys.ANYDATA.ConvertVarchar2(p_dayend_date);
l_result := sys.UTL_DBWS.invoke (
call_handle => l_call,
input_params => l_input_params);
sys.UTL_DBWS.release_call (call_handle => l_call);
sys.UTL_DBWS.release_service (service_handle => l_service);
EXCEPTION WHEN OTHERS THEN
PKG_IRIS_SH.PRC_ins_spm('WEB','9999', 'Web Service', TO_DATE(p_dayend_date ,'DD/MM/YYYY HH24:MI:SS') ,'Error Description:' ||SQLERRM , '0');
End ;
Can any body advise on this issue.
thanks,
Ram

Cross posted to APEX support forum, PLEASE CLOSE THIS THREAD..
Thank you,
Tony Miller
Webster, TX
Never Surrender Dreams!
JMS
If this question is answered, please mark the thread as closed and assign points where earned..

Performance issue on calling the API queryFolders to retrive the subfolders

Hi, We are using 10.3.0 collaboration server.When we make an API call to get the folders under a parent folder, the API is taking more than 2 minutes to return. When we did research, found out that the query collaboration server is executing taking time to return the results from the collaboration database. I am copying the query below. The performance issue coming when the number of sub folders is more than 1400. I did few experiments on this query. When I remove the sort clause, the query is running too fast. If i keep the sort clause and remove the sub query in select clause , query is running fast. But the combination is taking too much time i.e. 80 seconds. Please share your ideas on how to increase the performance of this API call. We are planning to upgrade to another version of collab but not sure whether it would fix the problem. Highly appreciated your valuable suggestions. Let me know if you need more information.
SELECT OBJ.OBJECTID, OBJ.TID, OBJ.CREATED, OBJ.MODIFIED, OBJ.NAME,
OBJ.DESCRIPTION, OBJ.OWNERID, OBJ.PROJECTID, OBJ.PARENTID,
OBJ.OWNERNAME, OBJ.PARENTPATH, OBJ.PARENTIDPATH, OBJ.DEPTH,
OBJ.MODERATED, OBJ.CRAWLERACCESSIBLE,(select DISTINCT(1) FROM
CSFOLDERS where(exists(select * from csFolders subFolders where
subFolders.parentID=OBJ.objectid))) as HASVISSUBFOLD FROM CSFOLDERS
OBJ WHERE EXISTS (SELECT * FROM CSROLEMEMBERSHIP RM,CSFOLDERSECURITY
OS,PLUMTREE.PTVGROUPMEMBERSHIP GRP WHERE OS.ACCESSLEVEL > 0 AND
GRP.USERID=:v0 AND GRP.GROUPID = RM.PTGROUPID AND RM.OBJECTID =
OS.ROLEID AND OS.OBJECTID = OBJ.OBJECTID) AND obj.TID = 3 AND
(obj.parentID=:v1) ORDER BY LOWER(obj.NAME) ASC

hi reddy,
before applying for all entries on mseg
1.sort t_mard by matnr werks and lgort
2. if not t_mard is initial.
select mblnr mjahr etc...
endif.
these are the main precautions u need to look into before you are using second select.
entires from bseg, mseg will consume lot of time.
so the best if only the fields u r showing are required then take it from the respective table like
ebeln ebelp from ekpo
like that .
so the alternative will be this but this will be the best option if u know only these fields are to be selected .
hope u r getting my point.
yeah swagatika also said correctly that u can go for indexing so the Retrieval is faster.
one more option is construct a view/structure and proceed .
regards,
TVK.
Message was edited by: Vikky

ResourceBundle throws MissingResourceException when calling Java from C

All,
Our code calls Java from C application.
The C code create JVM using JNI_CreateJavaVM. Due to internal design issues we don't
create the JVM with a "-classpath" option.
After creating the JVM, we create a ClassLoader (URLClassLoader), with a URL list as it's classpath.
The last step in the C code is to load a Java class using the C reference to URLClassLoader.loadClass
(I will refer to this class as classA), and than call one of it's static methods.
Please notice that classA is not instantiated, we call a static method.
At this point we are moving to the Java code...
In the Java static method we call ResourceBundle.getBundle("resources.messages"),
the result is ResourceBundle throws MissingResourceException.
- The JAR hosting "resources/messeges.properties" (refer as jarA) is include in the
ClassLoader URL paths we used for loading the Java class.
- Using a debugger we can see that ClassLoader.getSystemClassLoader() don't contain
jarA in it's URL list.
- Using the debugger we see that classA.class.getClassLoader() does contain
jarA in it's URL list.
After some investigation we see that ResourceBundle.getBundle calles to ResourceBundle.getLoader:
private static ClassLoader getLoader() {
        Class[] stack = getClassContext();
        /* Magic number 2 identifies our caller's caller */
        Class c = stack[2];
        ClassLoader cl = (c == null) ? null : c.getClassLoader();
        if (cl == null) {
            cl = ClassLoader.getSystemClassLoader();
        return cl;
    } getClassContext() - native method, this can explain the why the above work fine for Windows.
I suspect the problem raised in the above code. Since I couldn't get a debug
version of rt.jar for Solaris, I need some help.
Our environment is Solaris 8, Java 1.5.
Please note that the same code works fine on WindowsXP with the same Java version!
Thanks,
Avi

I was thinking JVM should the classpath when it is invoked. But obviously it is not and we need to assign the classpath ourselves when invoking from native side. Thanks for pointing that out. For other people's benefit the following is the code piece I for initializing JVM.
char* str1;
          char* str2;
          char* classpath;
          str1 = "-Djava.class.path=";
          str2 = std::getenv("CLASSPATH");
          classpath = (char*)malloc(strlen(str1) + strlen(str2) + 1);
          strcpy(classpath, str1);
          strcat(classpath, str2);
          if (str2 == NULL)     {
               printf("CLASSPATH environment variable is not defined.");
               exit(1);
          m_VmArgs.version = JNI_VERSION_1_4;
          m_options[0].optionString= classpath;
          m_VmArgs.options = m_options;
          m_VmArgs.nOptions = 1;
          m_VmArgs.ignoreUnrecognized = JNI_TRUE;
          JNI_GetDefaultJavaVMInitArgs(&m_VmArgs);
          /* Create the Java VM */
          m_Res = JNI_CreateJavaVM(&m_Jvm, (void**)&m_Env, &m_VmArgs);
          if (m_Res < 0) {
               printf("Can't create JVM\n");
               exit(1);
thanks,
chancellor

Performance issue when calling java module

Similar Messages

Maybe you are looking for